Predicting Rainfall for Farming in the Bantul Region Using an Artificial Neural Network

¹ Magister of Civil Engineering, Universitas Muhammadiyah Yogyakarta, Yogyakarta 55183, Indonesia
² Department of Information Technology, Universitas Muhammadiyah Yogyakarta, Yogyakarta 55183, Indonesia

^* Corresponding author: riyadi@umy.ac.id

Abstract

Climatic conditions of the rainy season, such as the clear difference between the rainy and dry seasons, greatly affect the meteorological characteristics, especially the temperature and rainfall in the territory of Indonesia. To maximize the availability of water and variations in rainfall for plant growth and development, plant cultivation requires a proper approach. The method used in this study is Artificial Neural Network which is implemented with the help of Matlab software version 2019 with nntools. This method is used to predict rainfall in the Bantul area. In this study, the data used were rainfall, minimum temperature, maximum temperature, average temperature, wind speed, humidity, and air pressure. This data is processed using Artificial Neural Networks to accurately predict rainfall in the region. The test results show that the comparison of the actual data results of rainfall prediction using the Levenberg Marquart algorithm with 1,080 training data of 80% data composition, validation data 10 and test data 10 with layer 4 size with layer 10 hidden neural produces predictions with a good level of accuracy and obtains a value of R = 0.900.